Metal oxide photoresist options for high-NA EUV patterning applications

High-NA EUV lithography represents the most significant step change in optical patterning since the introduction of immersion lithography. By increasing the numerical aperture from 0.33 to 0.55, high-NA EUV tools reduce the minimum half-pitch achievable in a single exposure to below 8 nm, enabling the feature densities required for logic nodes beyond 2 nm. That capability, however, places extraordinary demands on every layer in the patterning stack, and photoresist sits at the centre of those demands. Selecting the right resist chemistry, particularly among emerging metal oxide photoresist options, is one of the most consequential materials decisions process integration teams face as they qualify high-NA processes.

Metal oxide photoresists have attracted sustained attention from the semiconductor industry precisely because their fundamental material properties align well with the physical constraints high-NA EUV imposes. As a lithography materials supplier working at the boundary of next-generation patterning, understanding why organic resist chemistries struggle and what metal oxide systems offer in their place is essential groundwork for any evaluation programme.

The patterning demands driving high-NA EUV adoption

The transition to high-NA EUV is driven by the need to print sub-10 nm half-pitch features with sufficient process margin for high-volume manufacturing. At these dimensions, the aerial image delivered by a 0.55 NA optical column is sharper than anything achievable with 0.33 NA tools, but the depth of focus is correspondingly reduced. The usable focus window shrinks to a fraction of what process engineers managed under conventional EUV, compressing the tolerance for resist thickness, topography variation and underlayer interactions simultaneously.

Overlay and edge placement error budgets tighten in direct proportion to the reduction in feature pitch. At sub-8 nm half-pitch, an edge placement error of even a few angstroms can translate into yield-limiting shorts or opens. This means the resist must not only resolve fine features but must do so with minimal stochastic variation, because line-width roughness and local critical dimension uniformity become first-order yield drivers rather than secondary concerns.

Why organic chemically amplified resists struggle at high-NA

Chemically amplified resists, the workhorses of EUV lithography since its introduction, rely on acid-catalysed deprotection reactions triggered by photoacid generators distributed through a polymer matrix. The amplification mechanism is efficient in terms of photon utilisation, but it introduces a fundamental spatial blurring: acid diffusion during the post-exposure bake spreads the latent image beyond the boundaries of the absorbed photon distribution. At feature pitches above 20 nm, this diffusion is manageable. Below 10 nm, it erodes the contrast needed to resolve the aerial image faithfully.

The polymer backbone of organic resists also sets a practical lower limit on resolution through its own physical dimensions. High-molecular-weight polymer chains occupy volumes that are not negligible relative to sub-10 nm features, contributing to roughness at pattern edges. Reducing molecular weight to address this trades resolution improvement for sensitivity loss and etch durability reduction, creating a three-way constraint between resolution, sensitivity and line-edge roughness that organic chemistries have found difficult to escape at high-NA dimensions.

What makes metal oxide photoresists suited to high-NA EUV

Metal oxide photoresists replace the organic polymer matrix with inorganic or hybrid inorganic-organic networks built around metal centres, most commonly hafnium, tin or zirconium oxo-clusters. These clusters absorb EUV photons directly and with a high cross-section, because metal atoms present a far greater photoabsorption probability per unit volume than the carbon and hydrogen that dominate organic resist matrices. Greater EUV absorption per unit volume means more photochemical events per absorbed dose, which directly improves sensitivity without the diffusion penalty of chemical amplification.

The reaction mechanism in metal oxide resists is fundamentally different from chemically amplified systems. EUV-induced photolysis generates reactive species that cause condensation and crosslinking within the metal-oxo network, converting soluble clusters into an insoluble matrix in exposed regions. Because the crosslinking is localised at the molecular scale of the cluster network rather than spread by diffusion, the latent image retains the sharpness of the aerial image more faithfully. This translates into lower line-edge roughness and better local critical dimension uniformity at the feature sizes high-NA tools are designed to print.

The inorganic composition of metal oxide resists also confers inherently higher etch selectivity relative to underlying silicon or dielectric layers. This is a practical advantage in pattern transfer: a thinner resist film can be used without sacrificing the etch budget needed to open a hard mask, which in turn reduces the aspect ratio of the resist feature and improves its mechanical stability during development and rinse steps.

Key performance parameters for high-NA resist evaluation

Evaluating a metal oxide photoresist for high-NA EUV qualification involves a structured set of performance metrics that span optical, chemical and mechanical domains. No single parameter determines suitability; the evaluation must address the full set simultaneously.

Resolution and stochastic performance

Resolution is assessed by printing dense line-space arrays and contact hole arrays at the target pitch and measuring the minimum pitch at which features resolve without bridging or collapse. Stochastic performance is quantified through local critical dimension uniformity measurements across large sample populations and through defect density counts at the target dose. Metal oxide resists generally show improved stochastic performance at low doses relative to organic alternatives, but the specific cluster chemistry and formulation determine the magnitude of that advantage.

Sensitivity and dose-to-size

Dose-to-size, the exposure energy required to print a feature at its nominal critical dimension, must be low enough to support the throughput economics of high-NA tools. Because high-NA scanners operate with a smaller field size than 0.33 NA systems, throughput is already reduced relative to conventional EUV, making resist sensitivity a commercially significant parameter. Metal oxide resists with high EUV absorption cross-sections can achieve competitive dose-to-size values, though formulation optimisation for a specific underlayer and process stack is typically required to reach production targets.

Line-edge roughness and roughness power spectral density

Line-edge roughness is reported as a three-sigma value, but the spatial frequency distribution of that roughness, captured in the power spectral density, matters as much as the aggregate number. High-frequency roughness components are partially smoothed by etch processes, while low-frequency components transfer directly to the patterned feature. Resist evaluation programmes for high-NA applications increasingly specify power spectral density targets alongside the conventional three-sigma metric.

Integration considerations for metal oxide resist processes

Metal oxide photoresists introduce process integration requirements that differ from those of organic chemically amplified systems. Underlayer design is particularly important. The underlayer must promote adhesion of the metal-oxo cluster network to the substrate, provide the refractive index contrast needed to optimise the standing wave profile within the thin resist film, and be compatible with the aqueous or solvent-based developer used to remove unexposed material. Silicon-based underlayer chemistries, including silicon anti-reflective coatings and spin-on hard masks, have shown strong compatibility with metal oxide resist processes and provide the etch selectivity needed for robust pattern transfer.

Developer selection is a second integration variable. Metal oxide resists can be processed with aqueous base developers or with organic solvent developers, and the choice influences both resolution and defect performance. Solvent development has demonstrated advantages in reducing pattern collapse at aggressive pitches by lowering the surface tension forces acting on high-aspect-ratio features during the drying step. The interaction between developer chemistry, underlayer surface energy and resist network density requires systematic optimisation for each process node target.

Thermal stability during post-exposure bake and any subsequent thermal processing steps must also be verified. Metal-oxo networks can continue to condense at elevated temperatures, potentially shifting the critical dimension of printed features if bake conditions are not tightly controlled. Process window characterisation across bake temperature and time is a standard part of metal oxide resist qualification.

Advancing metal oxide resist chemistry for next-node readiness

The development of metal oxide photoresists for high-NA EUV is an active field, with ongoing work directed at improving sensitivity, reducing metal contamination risk and extending the chemistry to accommodate the even finer features that future nodes will require. One direction involves tuning the organic ligand shell surrounding the metal-oxo core to adjust solubility contrast between exposed and unexposed regions, enabling sharper development discrimination without sacrificing film uniformity. Another involves co-formulation strategies that combine multiple metal centres to optimise both absorption cross-section and network mechanical properties.

The introduction of high-NA EUV into high-volume manufacturing also raises questions about metal contamination management within the fab environment. Metal oxide resists deposit trace quantities of metal on wafer surfaces and on tool components, and process integration teams must establish cleaning and monitoring protocols to prevent cross-contamination. This is an area where collaboration between resist suppliers, equipment manufacturers and process integration engineers is essential, because the acceptable contamination budget is defined by the device architecture rather than by the resist chemistry alone.

PiBond’s work in inorganic EUV photoresist development, supported by a technology licence and strategic collaboration with the Paul Scherrer Institute, addresses the next-node requirements of high-NA EUV patterning through proprietary metal oxide and silicon-based resist chemistries. The company’s in-house monomer synthesis and ultra-pure cleanroom production capabilities allow resist formulations to be developed and scaled within a single organisation, reducing the variability that can arise when chemistry development and manufacturing are separated across the supply chain. For process integration teams evaluating resist options for high-NA qualification, engaging with a dedicated advanced patterning materials partner early in the development cycle is the most reliable path to aligning resist chemistry with the specific requirements of the target process stack.

As high-NA EUV moves from early adoption into broader production deployment through 2026 and beyond, metal oxide photoresist chemistry will continue to evolve in step with the patterning requirements it must meet. The fundamental material advantages these systems offer, in absorption efficiency, resolution fidelity and etch selectivity, position them as the most credible resist platform for the sub-8 nm half-pitch era, provided that integration challenges are addressed through systematic process development and close collaboration across the materials supply chain.

Related Articles

This content was generated with the help of AI and it may contain mistakes

‹See all news